About fifty percent of prematurely born infants develop learning/behavior problems and school failure. The study will test the primary hypotheses, that preterm infants (PT) randomized to developmental care in the Newborn Intensive Care Unit (NICU) will be superior in cognitive performance at school age when compared to their peers, who did not receive the intervention. Fifty-two [26 control (C);26 experimental (E)] infants born at a gestational age (GA) of 28-33 week (w) will be studied at 8 years(y) corrected age (CA). A sample of 26 fullterm-born (FT) healthy 8y-olds will provide an additional control group for descriptive purposes. Three secondary hypotheses will be tested: The PT-E group will be superior to the PT-C, 1) in academic performance;2) in long distance cortical connectivity of frontal and occipital brain systems (EEC coherence), and 3) in frontal cortical white matter volumes (MRI). The FT group will be superior to both PT groups. Child physical and psychosocial functioning will be explored descriptively. The primary independent variable will be treatment group (PT-control, PT-experimental). Additionally tested will be GA, growth percentile (weight and head circumference) at birth, parent socioeconomic status, and parent cognitive function, all expected to influence outcome. The PT sample will be described by previously collected medical and demographic measures at birth, and by behavior and brain function at 2wCA and 9mCA. For testing the primary (school age cognitive function) and the two secondary dependent variables (frontal coherence and frontal white matter volume), stepwise multiple linear regression analysis will be used;the five independent variables will be entered into the models. Main effects and two and three-way interaction terms will be evaluated with goodness-of-fit assessed using adjusted R-squared. A sample size of 26 subjects per PT group will provide 80% statistical power (beta=0.20;alpha=0.05, 2-tailed) to detect a 12 point or higher mean difference in the cognitive measure, assuming a pooled standard deviation of 15 points (effect size sigma = 12/15 = 0.8) based on a two-sample Student t-test. The study will provide, for the first time, evidence at school age of the effectiveness of a much needed, low cost, early behavioral intervention. The expected reduction in learning and school failure for the increasing numbers of preterm-born children in the nation's schools, would give the study and the intervention, significant national public health importance.,